1. Introduction

For half a century, the presence of 0.1 mV (100 uV) of horizontal or downsloping ST segment depression (1 mm at standard gain) has been the empirical ECG criterion for identification of coronary disease. This traditional criterion performs poorly, and a fixed threshold partition has a number of theoretical and clinical limitations as a marker for the presence and extent of coronary artery disease. In practice, the ST segment in an individual patient with effort-induced myocardial ischemia fall progressively during the course of diagnostic exercise testing. There is little reason to believe that coronary disease exists only when depression crosses the 1 mm level and not before. Progressive repolarization abnormality occurs because observed ST depression throughout the test depends not only on the extent of underlying coronary artery obstruction, but also on the metabolic severity of myocardial ischemia as it increases with ongoing cardiac work. What is needed is a way to adjust changing ST segment depression for this additional factor.

2. Background

Studies in England and in Hungary in the early 1980's suggested that relating the magnitude of ST segment depression change to heart rate change during peak effort by linear regression could improve the performance of the exercise ECG [Elamin et al., 1980; Berenyi et al., 1984]. These heart rate adjusted methods quantify the electrocardiographic "stress-strain" relationship between induced ischemia and the balance of energy demand of the heart and coronary blood flow during exercise. During the past 15 years, we have examined the value and limitations of heart rate adjustment of ST depression for the evaluation of coronary disease during treadmill testing [Kligfield et al., 1989; Okin and Kligfield, 1995a]. We have also examined the methodologic factors that govern the performance of these methods. In a number of clinically relevant populations, these methods can increase the sensitivity of the exercise ECG for the detection of coronary obstruction. Heart rate adjustment of ST segment depression also can improve the value of the exercise test for the identification of anatomically and functionally extensive disease. Moreover, these methods have been shown to significantly increase the predictive value of the easily accessible exercise ECG for coronary events and for cardiovascular mortality in asymptomatic men and women in the Framingham study population and for higher risk asymptomatic men in the MRFIT population. The present summary will focus only on heart rate adjustment of exercise phase ST segment data, but extension of similar principles to the analysis of recovery phase ST segment changes as a function of heart rate provides useful complementary data [Okin and Kligfield, 1995a; Lehtinen et al., 1996]. Of course, these are just better tests, not perfect tests.

3. Physiologic Basis for Heart Rate Adjustment of ST Segment Depression

A number of interpretive problems arise from traditional dependence on a fixed amount of ST segment depression as a diagnostic criterion for the detection of coronary artery disease during exercise. One may ask why 1 mm of ST depression in a patient exercising from a resting heart rate of 60 to a peak rate of 180 should be defined by empiric criteria as evidence of ischemia, while 0.5 mm of ST depression in another patient reaching a peak rate of only 90 is not? Surely, the amount of observed ST depression at any point in the exercise test must be related to exercise workload to evaluate the extent of coronary disease, or we are left with a difficult paradox. If a patient with 1 mm of ST depression during exercise continues to work harder and ST depression reaches 2 mm, is the extent of coronary disease increasing? If the same patient were stopped earlier in exercise, while ST depression was only 0.5 mm, would underlying coronary disease be any less?

Within this context, heart rate adjustment of measured ST segment depression is a physiologically sensible approach to quantification of the exercise ECG. It has long been established that changing heart rate during higher levels of exercise is directly related to changing myocardial oxygen consumption. When ST depression is plotted against heart rate during peak exercise-induced ischemia in patients with coronary disease, there is usually a close linear relation between the two variables. While additional ST depression occurs with increasing heart rates as workloads increase, the rate of change of this linear relationship remains constant in most patients with myocardial ischemia. Moreover, the linear rate of change of the ST segment- heart rate relationship can be generally related to the anatomic extent of coronary obstruction. These relationships can be described physiologically and modeled mathematically [Okin and Kligfield, 1995a].

3.1. Physiologic Description

From a physiologic point of view, heart rate adjustment of ST segment depression normalizes the increasing magnitude of apparent ischemia during exercise (as measured by changing ST segment depression) for the corresponding increasing myocardial workload (as measured by changing heart rate) that leads to ischemia in the presence of coronary disease. Consideration of factors that influence the magnitude of changing ST segment depression throughout exercise indicates that:

ST depression Extent of disease x Exercise workload

(1)

Since higher levels of changing exercise workload are directly proportional to heart rate, this can be rearranged as:

Extent of disease ST depression/ heart rate

(2)

From this it can be seen that it is the slope of the linear ST segment-heart rate relation that reflects the extent of underlying coronary disease, not the variable magnitude of ST segment depression alone. This relationship can be quantified by linear regression of data from the final stages of exercise, when ischemia is progressive in severity, as:

Extent of disease = d(STD)/d(HR)

(3)

where
STD = ST segment depression
HR = Heart rate

Alternatively, the overall average relation can more simply be quantified as:

Extent of disease = .

(4)

3.2. Solid Angle Analysis

These principles can also be examined within a simplified solid angle theory model that contains spatial and non-spatial terms [Okin and Kligfield, 1994]. In this framework,

(5)

where

dSTD = The magnitude of ST segment depression
= The solid angle subtending the ischemic boundary (the spatial term)
d(V) = The voltage change across the ischemic boundary (the non-spatial term)

(6)

so that

(7)

where c = constant

Division of Eq. 5 by d(HR) indicates that:

and substitution from Eq. 7 reveals that

Accordingly, the rate of change of ST segment depression with respect to heart rate during exercise can be theoretically and experimentally linked to the spatial extent of ischemia, separate from the metabolic severity of ischemia that varies with exercise workload. As a result, the magnitude of the slope relating ST depression to changing heart rate during peak exercise-induced myocardial ischemia should increase with the anatomic severity of coronary artery disease, whereas measured ST depression alone also should vary continuously with the exercise workload achieved. These observations provide a basis for understanding and interpreting the heart rate adjusted indices as useful continuous variables that reflect both the presence or absence of coronary obstruction and its anatomic severity.

4. ST/HR Slope and Simple ST/HR Index

Methods of heart rate adjustment of the exercise ECG include the fairly complex, linear regression-based ST segment/heart rate (ST/HR) slope and the simpler ST/HR index [Okin and Kligfield, 1995]. Both methods report the rate of change of ST depression with respect to changing exercise heart rate in units of microvolts/beat/minute (uV/bpm). The ST/HR slope corresponds to Eq. 3. It seeks to quantify the maximum rate of ST segment change with respect to heart rate during the period of active ischemia that occurs at the end of exercise. Details of this calculation are available in previous reports. The ST/HR index corresponds to Eq. 4. It is a practical approximation of the ST/HR slope that results from simple division of the overall maximal change in ST segment depression by the corresponding overall change in heart rate during exercise [Detrano et al., 1986, Kligfield et al., 1989]. As a result, the ST/HR index represents the average rate of change of ST depression over the entire course of exercise. The simple ST/HR index calculation includes the period of early exercise in which ischemia is not present, in contrast to the maximal rate of change at peak exercise that is described by the regression-based ST/HR slope. Accordingly, values calculated for the ST/HR index are proportionately lower than corresponding values calculated for the ST/HR slope. The difference in these measures is illustrated in Fig. 1.

Figure 1. Derivation of the ST/HR slope by linear regression of end-exercise data and the ST/HR index from control and peak exercise data points only. The ST/HR index is quite simple to calculate, but it generally underestimates the ST/HR slope and it is strongly dependent on the precision of measurement of the peak ST segment depression.

5. Performance of Heart Rate Adjusted Measures of ST Depression

In clinically relevant populations, the ST/HR slope can increase the sensitivity of the standard exercise test for the detection of coronary disease, with high specificity, and it can also identify patients with anatomically and functionally severe disease [Okin et al., 1988; Kligfield et al., 1989]. Values for the ST/HR slope <2.4 microvolts/beat/minute (uV/bpm) have been found in approximately 95% of normal men and women, while values of 2.4 uV/bpm or more occur in approximately 90% of patients with stable angina pectoris and with catheterization proved coronary artery disease. In practice, the ST/HR slope is also useful for assessing the severity of coronary disease, because its magnitude increases with the extent of disease. Patients with anatomically or functionally extensive coronary obstruction, such as those with left main or proximal three vessel coronary disease, generally have ST/HR slope values of 6.0 uV/bpm or more.

Sensitivity and specificity of the ST/HR index for the detection of coronary disease approach that of the more complex ST/HR slope, but the simple index does not perform as well as the regression-based method for the assessment of the anatomic and functional severity of disease [Okin and Kligfield, 1995a]. Values for the simple ST/HR index have been found to be <1.6 uV/bpm in approximately 95% of normal men and women, and 1.6 uV/bpm or more in approximately 90% of patients with coronary artery disease. Patients with left main or proximal three vessel coronary disease often have ST/HR index values of 3.3 uV/bpm or more.

Peak exercise horizontal or downsloping ST depression of 1.0 mm identified 62% of 337 patients with coronary artery disease [Okin and Kligfield, 1995b], including 67% of the 246 men but only 51% of the 91 women (p<0.01). Reduced standard test sensitivity in women can be related at least in part to the smaller magnitude of ST segment depression that occurs at lower levels of peak effort tolerance. At matched specificity of 96%, simple heart rate adjustment of end-exercise phase ST segment depression by the ST/HR index improved the sensitivity of the test to 91% in men and to 85% in women. Among asymptomatic but higher risk men in the MRFIT study, the relative risk of coronary death among those with an abnormal ST/HR index was 4.0 [Okin et al., 1996]. Among asymptomatic adults in the Framingham Offspring study, the relative risk of coronary events during a four year followup period was increased to 3.1by means of the ST/HR index [Okin et al., 1991]. While relative risk was 2.6 in the 1521 men, it was 5.4 in the 1647 women. Thus, the improved performance of the simple ST/HR index during exercise testing is applicable and particularly important in women.

6. Value and Limitations of Heart Rate Adjusted ST Depression

The improved sensitivity of the ST/HR slope and the ST/HR index for the detection of coronary disease can be related to several factors. These include correct classification of patients with "equivocal" standard test responses, identification of patients with only one- and two-vessel disease who are often missed by the standard test, and appropriate workload adjustment of patients with "subthreshold" ischemic depression [Kligfield et al., 1989, Okin and Kligfield, 1995]. However, these are hardly perfect tests, and these methods have not performed well in all studies [Lachterman et al., 1990]. Performance will vary with population selection and workup bias [Okin and Kligfield, 1995; Morise, 1997], as exemplified by the significantly lower specificity in catheterized patients with normal coronary arteries. False positive test responses occur with the ST/HR slope and the simple ST/HR index in patients with hypertrophic, myopathic, and valvular disease, and in patients with bundle branch block. False negative responses are common after recent Q wave infarction. The effects of drug and hormones on the heart rate adjusted measures require further study. While the ST/HR slope is applicable and comparable in value in patients taking beta-blocking drugs, false negative responses can occur in patients whose exercise tolerance is so low that data are insufficient for regression.

References